Content uploaded by Filip Kostelka
Author content
All content in this area was uploaded by Filip Kostelka on Apr 22, 2018
Content may be subject to copyright.
1
The Chicken and Egg Question: Satisfaction with Democracy and Voter Turnout
Filip Kostelka1 & André Blais2
Abstract: Political scientists, analysts and journalists alike have for long believed that the
degree of satisfaction with the functioning of democracy determines voter turnout. In this
article, we use survey data from 24 panel studies and demonstrate that the causal relationship
is actually reversed: voter turnout affects satisfaction with democracy and not the other way
around. We also show that this reversed relationship is conditioned by election type, electoral
system, and election outcomes. These finding are important since: a) They question
conventional wisdom and a large body of scientific literature; b) They invite a more nuanced
approach in the study of the relationship between evaluations of regime performance and
political participation; c) They underline the central role of elections in shaping citizens’
perception of the democratic process.
Note: This is a manuscript accepted for publication in PS: Political Science and Politics
(forthcoming probably in April 2018). Following the journal’s style, the manuscript uses
endnotes. Suggested citation:
Kostelka, Filip and André Blais, “The Chicken and Egg Question: Satisfaction with
Democracy and Voter Turnout”, PS: Political Science and Politics, forthcoming.
1 Filip Kostelka is a postdoctoral fellow at the Research Chair in Electoral Studies, University
of Montreal, Canada, and associate researcher at the Centre d’études européennes, Sciences
Po, Paris, France. He coordinates the Making Electoral Democracy Work project
(http://electoraldemocracy.com) and can be reached at filip.kostelka@umontreal.ca.
2 André Blais is professor in the department of political science and holds a Research Chair
in Electoral Studies at the University of Montreal, Canada. He leads the Making Electoral
Democracy Work project (http://electoraldemocracy.com) and can be reached at
andre.blais@umontreal.ca.
2
Introduction
“Does a low election turnout indicate voters are disillusioned or content?” asked one
of the world’s oldest newspaper The Herald in the run-up to the 2003 election to the Scottish
Parliament (The Herald 2003, 12). In its article, the Glasgow-based broadsheet comes to the
conclusion that some abstainers may “register their disillusion” while others may be “content
with the way things are”. Interestingly, as we review below, this intuition reflects the current
state of the scientific debate on the impact of citizen’s satisfaction on voter turnout.
Nevertheless, the title is illustrative of yet another phenomenon. More broadly, it shows how
pundits usually conceptualize the relationship between satisfaction with the way democracy
works and voter turnout: the former as the cause and the latter as the effect. However, such a
view is incomplete since there are theoretical reasons and empirical findings for considering
the presence of a reversed relationship. In this article, we use data from 24 panel studies and
debunk the conventional wisdom. We find no evidence of the effect of democratic satisfaction
on voter turnout but robust support for the reversed relationship. In other words, voter turnout
tends to affect satisfaction and not the other way around. In national and subnational elections,
the effect is positive and it is stronger in majoritarian electoral systems and among voters who
think that their preferred party won the election. Conversely, in supranational elections, the
effect is negative and affects the whole electorate.
Hypotheses
Satisfaction with democracy is best understood as an indicator of regime performance,
situated between more diffuse support for political community and regime principles and more
specific support for regime institutions and political actors (Norris 1999, Linde and Ekman
2003, Norris 2011, Linde 2012). As regards its impact on voter turnout, the political science
literature has been divided in two camps (Pacek et al. 2009; Ezrow and Xesonakis 2014). On
the one hand, especially in earlier works, nonvoting is often seen as a sign of satisfaction with
3
the current state of affairs (e.g. Tingsten 1930, Wilson 1936, Lipset 1981). On the other hand,
especially in the more recent literature, “dissatisfaction with democratic performance is
usually regarded at least implicitly, as an important cause of civic disengagement” (Norris
2011; see also Kostadinova 2003). Both perspectives agree on the direction of causality and
assume that one’s perception of the overall functioning of the political system motivates the
decision to vote or abstain. Political scientists, thus, frequently use satisfaction with
democracy to predict voter turnout (e.g. Grönlund and Setälä 2007; Flickinger and Studlar
2007, Hadjar and Beck 2010).
Hypothesis 1a: Satisfaction with democracy decreases voter turnout.
Hypothesis 1b: Satisfaction with democracy increases voter turnout.
Most empirical studies employ post-election measures, or study macro-level data
using measurements from non-electoral surveys. The two methods are problematic. First, the
macro-data approach can provide only indirect evidence of the causal mechanism. Secondly,
the post-election type of measurement implicitly presupposes that the election itself has not
altered citizens’ level of democratic satisfaction. Such presupposition is questionable.
Elections are at the heart of contemporary democracy (Dahl 1971; Huntington 1991).
They are the principal means of changing the political course of a democratic polity. It is
through elections that an unpopular incumbent can be voted out of office and replaced by a
more popular alternative. They represent the most tangible embodiment of the democratic
principle to which citizens are regularly and systematically exposed. It would be surprising if
elections had no effect on citizens’ view about how democracy works in their country. We
hypothesize that, in most circumstances, elections boost satisfaction with democracy,
especially among voters who play the democratic game.
Hypothesis 2: Electoral participation increases satisfaction with democracy.
However, the impact of electoral participation is likely to vary across contexts and
individuals. With regard to the former, an increase in satisfaction can be expected only as long
4
as there is a clear link between election outcomes and government composition. A
quintessential example of elections where such a link is tenuous are supranational elections to
the European Parliament. Although they are run essentially as national contests on national
issues (Reif and Schmitt 1980), a party winning the national vote may end up in the losing
camp at the supranational level. Furthermore, the European Union’s institutional structure and
operating mechanisms further hinder accountability (Follesdal and Hix 2006). Consequently,
the positive effect of elections on satisfaction should be weak or inexistent in EP elections. On
a more general level, following Aarts and Thomassen 2008, we expect that elections under
majoritarian electoral rules allow for more accountability and, therefore, boost satisfaction
with democracy more strongly than elections using proportional representation.1
Hypothesis 3: The positive effect of electoral participation on satisfaction with
democracy is weak or inexistent in European Parliament Elections.
Hypothesis 4: The positive effect of electoral participation on satisfaction is
stronger in majoritarian electoral systems
As for individual-level variations, it is obvious that election outcomes are not equally
liked or disliked by all voters. Citizens’ preferences presumably condition election-related
change in satisfaction. Winners, i.e. those who prefer the party (or parties) that won the
election, are likely to see their satisfaction increase much more than losers, i.e. those who
prefer the losing alternatives.
Hypothesis 5: Satisfaction with democracy increases the most among election
winners.
The impact of elections and election outcomes on mass political attitudes has been
studied empirically by a vibrant strand of political science literature. These studies have
generally found that those who participate in an election consider the outcome of the election
as more legitimate than abstainers (Nadeau and Blais 1993) and that winners become more
satisfied than losers (Anderson et al. 2005; Henderson 2008; Singh 2014; Campbell 2015;
5
Singh and Thorton 2016). Although these studies provide valuable insights, they suffer from
the same limitation as the aforementioned research since they are based on the same kind of
data – post-electoral surveys. They consequently do not allow researchers to control for pre-
election attitudes. As for panel studies, they are rare, usually cover a single election and often
study different, typically more specific, types of political support. They have found that voting
in elections increases external efficacy (Ginsberg and Weissberg 1978) and perceptions of
legislators’ responsiveness (Clarke and Kornberg 1992). Those effects tend to be the strongest
among election winners, who also become more trustful of government (Anderson and
LoTempio 2002) and less cynical about political parties (Banducci and Karp 2003). In terms
of satisfaction with democracy, a positive effect has been detected in the 1999 legislative
election in New Zealand (Banducci and Karp 2003) in the 2010 North Rhine-Westphalia
election (Singh et al. 2012) and in the French presidential election of 2012 (Beaudonnet et al.
2014). Blais and Gélineau 2007 found a positive effect on voters in general and winners in
particular in the Canadian election of 1997. Finally, Blais et al. 2015 found that satisfaction
increased among those who voted for parties that gained greater shares of votes, seats and
cabinet portfolios. All these findings reveal that the use of post-electoral measures of
satisfaction can be a risky strategy since causality may go in the opposite direction. We thus
proceed to a systematic examination of the relationship between satisfaction with democracy
and voter turnout.
Data
To test the five hypotheses, we use individual-level survey data from the Making
Electoral Democracy Work project (MEDW, Blais 2010). The project studies national,
subnational, and supranational elections held between 2010 and 2015 in five countries
(Canada, France, Germany, Spain and Switzerland).2 Our dataset consists of 24 region-
election samples that cover 17 elections held between 2010 and 2015 (five national, nine
subnational and three supranational, see the Electronic Appendix). These data are particularly
suitable for disentangling the satisfaction-turnout nexus since it has a two-wave panel structure
6
and provides both a pre-election and post-election measure of satisfaction with democracy.
Respondents were asked, generally one week before and week after the given election, to
indicate their degree of satisfaction with democracy at the election-related level of governance.
They were proposed a 0-10 scale where 0 means “not satisfied at all” and 10 “very satisfied”.3
We divide the scale by ten so that it varies between 0 and 1.
The control variables are the classic predictors of electoral participation commonly
used in political science research such as age, education and political interest (Blais 2007, see
the Electronic Appendix for descriptive statistics). It should be noted that all the control
variables were measured in the pre-election surveys while the main variable of interest –
reported electoral participation – was measured after the election. Hypothesis 3 and 4 are
tested by dummy variables that distinguish different types of elections (national, subnational
and supranational) and different electoral systems (majoritarian and proportional). As for
Hypothesis 5, we introduce two dummy variables that ascertain whether voters think the party
for which they had cast a vote won or lost (abstainers and don’t knows are the reference
category).4 Finally, all the of the following analyses include region (or election-region)
dummies.
Findings
In Table 1, we present the results of four analyses. Model 1 regresses electoral
participation on the pre-election and post-election measures of satisfaction with democracy
alone.5 In conformity with Hypothesis 2, the post-election measure is much more strongly
associated with voting than the pre-election one. At this stage, it cannot, however, be excluded
that pre-election satisfaction, the coefficient of which is also positive and statistically
significant, fosters voting too as Hypothesis 1b suggests. Therefore, in Model 2, we test the
pre-election measure alone while incorporating the control variables. The coefficient is now
even smaller than in model 1 and no longer meets the most lenient threshold of statistical
7
significance. This means that the weak positive effect of the pre-election measure in Model 1
was essentially a proxy for variables that are more closely related to voter turnout such as
education or interest in politics. People who are dissatisfied with democracy before the
election do not participate less because they are dissatisfied but because they are less educated
or uninterested in politics. In contrast, the effect of the post-election measurement in Model 3
is still substantial and meets the most rigorous threshold of statistical significance.
Furthermore, the findings from Models 2 and 3 are confirmed in Model 4 where the two
measures are tested jointly with the control variables. While the post-election measure remains
statistically significant, the pre-election measure changes sign and is insignificant. All the
models point in the same direction: conventional hypotheses (1a and 1b) are rejected while
the first alternative hypothesis (2) is supported.
Figure 1 graphically expresses the relationship between electoral participation and
post-election satisfaction. It displays the average marginal effects (AME) calculated from
Model 3. If post-election satisfaction was the cause and voting the effect, it could be
understood as the change in the probability of voting when respondents’ value on satisfaction
increases. If we did not have the pre-election measure, we would conclude that a shift from
no satisfaction to maximal satisfaction increases the probability of voting by approximately
five percentage points. This is a large effect, which would provide strong evidence for the
conventional interpretation (Hypothesis 1b). We know, however, that turnout is related to
post-election satisfaction and not to pre-election satisfaction, and thus that it is voting that
boosts satisfaction and not the other way around.
8
Table 1 Turnout and Satisfaction with Democracy
(1) (2) (3) (4)
No
controls
Pre &
Controls
Post &
Controls
Post &
Pre &
Controls
Pre
-
Election Satisfaction
0.28
*
0.19
-
0.11
(0.12)
(0.12)
(0.14)
Post
-
Election Satisfaction
0.90
***
0.44
***
0.51
***
(0.12)
(0.12)
(0.14)
Female
-
0.11
-
0.11
-
0.11
(0.06)
(0.06)
(0.06)
Age
0.02
***
0.02
***
0.02
***
(0.00)
(0.00)
(0.00)
Post
-
Secondary Education
0.31
***
0.30
***
0.30
***
(0.06)
(0.06)
(0.06)
Duty to Vote
1.18
***
1.17
***
1.17
***
(0.06)
(0.06)
(0.06)
Interest in Politics
1.14
***
1.11
***
1.12
***
(0.12)
(0.12)
(0.12)
Political Knowledge
1.26
***
1.25
***
1.25
***
(0.10)
(0.10)
(0.10)
Feel close to a Party
0.55
***
0.54
***
0.54
***
(0.07)
(0.07)
(0.07)
Election
-
Region Dummy Variables
Yes
Yes
Yes
Yes
Constant
0.71
***
-
1.52
***
-
1.66
***
-
1.64
***
(0.12)
(0.17)
(0.17)
(0.17)
Observations
19076
19076
19076
19076
Pseudo
R
2
0.04
0.18
0.18
0.18
Note: Logit regression coefficients with robust standard errors in parentheses. * p < 0.05, ** p < 0.01,
*** p < 0.001.
9
Figure 1 Predicted Probability to Vote and Satisfaction with Democracy
The impact of voter turnout on satisfaction with democracy is formally tested in Table
2. The dependent variable is the change in satisfaction with democracy between the pre-
election and post-election measurements. The main predictor of interest is turnout. To take
into account ceiling effects, we control for pre-election satisfaction. The result clearly
corroborates Hypothesis 2: having voted increases satisfaction with democracy by 0.02 (0.2
on the 0-10 scale) among voters when compared to abstainers (the reference category).
10
Table 2 Pre-Post Change in Satisfaction with Democracy and Turnout
B
Voted 0.02*** (0.01)
Pre-Election Satisfaction -0.43*** (0.01)
Female -0.01 (0.00)
Age -0.00 (0.00)
Post-secondary Education 0.01*** (0.00)
Interest in Politics 0.04*** (0.01)
Political Knowledge 0.03*** (0.01)
Feel Close to a Party 0.01** (0.00)
Election-Region Dummy Variables Yes
Constant 0.23*** (0.01)
N 19076
R2 0.25
Note: OLS Regression coefficients with robust standard errors in parentheses. * p < 0.05, ** p < 0.01,
*** p < 0.001
To test Hypotheses 3 and 4, we added election type and electoral system as predictors
in the model presented in Table 2. Each variable was also interacted with voting. We plot the
effect of these variables as average marginal effects in Figure 2 and 3.6. They express the
predicted change in satisfaction with democracy for voters and abstainers in different
institutional contexts.
In terms of election types (Figure 2), national and subnational elections increase
voters’ satisfaction by approximately 0.03 while exerting no effect on abstainers. Conversely,
supranational elections to the European Parliament depress satisfaction in the whole electorate
by practically the same amount. This validates and even exceeds the expectations of
Hypothesis 3. It suggests that EP elections expose the democratic imperfections of the
European Union and, instead of boosting democratic satisfaction, they decrease it.
11
Figure 2 Predicted Pre-Post Change in Satisfaction with Democracy by Election
Type
Given the peculiarity of supranational elections, in the remaining analyses, we focus
on national and subnational elections. Figure 3 displays the effects of electoral systems. In line
with Hypothesis 4, elections run under majoritarian rules increase satisfaction among voters
twice as strongly (0.04) as elections in proportional systems (0.02). Like in Figure 2,
abstainers’ level of satisfaction remains stable in both types of systems.
Figure 3 Predicted Pre-Post Change in Satisfaction with Democracy by
Electoral System
12
Finally, we examine the impact of perceptions of election outcomes. Figure 4 shows
that, as predicted by Hypothesis 5, the effect is much stronger among election winners. Those
who believe that their preferred party won the election see their satisfaction increase by 0.09
(0.9 on the 0-10 scale). Nevertheless, a substantially small but statistically significant increase
is observable also among election losers. This reveals that elections tend to legitimize
democracy among all voters and not only winning parties’ supporters.
Figure 4 Predicted Pre-Post Change in Satisfaction with Democracy among
Voters
Conclusion
This article questions the conventional view about the causal relationship between
satisfaction with democracy and electoral participation. What is thought by many to be the
cause is in reality the effect and vice versa. This finding has important implications.
First, the results of earlier studies, based on post-electoral surveys, need to be
reinterpreted. In light of our findings, those prior works do not testify to the impact of
13
satisfaction on voting but, instead, to the impact of voting on satisfaction. There is no evidence
that those who are less satisfied before the election vote less.
Second, our analysis contributes to the broad literature on the role of elections in
democracies. It provides robust support for the legitimizing effect of elections. We show that,
in national and subnational elections, especially those run under majoritarian rules, voting
increases the satisfaction of all voters – even those who lost the election – and that it strongly
boosts the satisfaction of those who believe that they won the election. In that respect, the
generalized boost in satisfaction with democracy resembles the honeymoon effect, which also
affects both election winners and losers (e.g. Brody 1991). Future studies should investigate
to what extent the two phenomena are related and whether the election-driven boost in
satisfaction with democracy is as temporary as the honeymoon.
There are certainly circumstances in which citizens can become less satisfied after an
election. In this study, we found that supranational elections decrease satisfaction both among
voters and abstainers. Such a negative effect can be expected in other contexts in which
elections produce outcomes that are seen as distorted, illegitimate or ineffective, and it may
affect especially those who voted for parties that fail to gain any representation (Blais et al.
2015). For instance, in the U.S. context, satisfaction may decrease among those voters whose
preferred presidential candidate loses the White House despite winning the popular vote like
the Democrats in 2000 and 2016 (see Craig et al. 2006). More generally, future research should
explore in greater detail how contextual factors such as disproportionality or party system
fragmentation condition the effect of elections on citizens’ satisfaction.
Third, our findings underline the pitfalls of using satisfaction with democracy
measured in post-electoral surveys to predict voting behaviour. These attitudes are likely to
be affected by the election itself. The causal arrow may go in the other direction: from voting
behaviour to attitudes.
14
Fourth, our results remind us that elections are a central democratic institution. In
retrospect, it seems obvious that elections affect how citizens perceive the functioning of
democratic regimes. This does not rule out the possibility that other factors such as political
scandals may produce strong shifts in satisfaction with democracy (Kumlin and Esaiasson
2012). The bottom line, however, is that elections matter and that, most of the time,
participating in an election makes citizens more satisfied with the way democracy works.
References
Aarts, Kees, and Jacques Thomassen. 2008. “Satisfaction with Democracy: Do
Institutions Matter?” Electoral Studies 27 (1): 5–18.
Anderson, Christopher J. et al. 2005. Losers’ Consent:Elections and Democratic
Legitimacy: Elections and Democratic Legitimacy. Oxford: Oxford University
Press.
Anderson, Christopher J., and Andrew J. LoTempio. 2002. “Winning, Losing and
Political Trust in America.” British Journal of Political Science 32 (2): 335–51.
Banducci, Susan A., and Jeffrey A. Karp. 2003. “How Elections Change the Way
Citizens View the Political System: Campaigns, Media Effects and Electoral
Outcomes in Comparative Perspective.” British Journal of Political Science 33
(3): 443–67.
Beaudonnet, Laurie, André Blais, Damien Bol, and Martial Foucault. 2014. “The
Impact of Election Outcomes on Satisfaction with Democracy under a Two-
Round System.” French Politics 12 (1): 22–35.
Blais, André. 2007. “Turnout in Elections.” In The Oxford Handbook of Political
Behavior, eds. Hans-Dieter Klingemann and Russell J Dalton. Oxford: Oxford
University Press, 621–35.
———. 2010. “Making Electoral Democracy Work.” Electoral Studies 29 (1): 169–
70.
Blais, André, and François Gélineau. 2007. “Winning, Losing and Satisfaction with
Democracy.” Political Studies 55 (2): 425–41.
Blais, André, Alexandre Morin-Chassé, and Shane P. Singh. 2015. “Election
Outcomes, Legislative Representation, and Satisfaction with Democracy.” Party
Politics: 1354068815583200.
15
Brody, Richard. 1991. Assessing the President: The Media, Elite Opinion, and Public
Support. Stanford, CA: Stanford University Press.
Campbell, Ross. 2015. “Winners, Losers and the Grand Coalition: Political
Satisfaction in the Federal Republic of Germany.” International Political
Science Review 36 (2): 168–84.
Clarke, Harold D., and Allan Kornberg. 1992. “Do National Elections Affect
Perceptions of MP Responsiveness? A Note on the Canadian Case.” Legislative
Studies Quarterly 17 (2): 183–204.
Craig, Stephen C., Michael D. Martinez, Jason Gainous, and James G. Kane. 2006.
“Winners, Losers, and Election Context: Voter Responses to the 2000
Presidential Election.” Political Research Quarterly 59 (4): 579–92.
Dahl, Robert Alan. 1971. Polyarchy: Participation and Opposition. New Haven, CT:
Yale University Press.
Ezrow, Lawrence, and Georgios Xezonakis. 2016. “Satisfaction with Democracy and
Voter Turnout A Temporal Perspective.” Party Politics 22 (1): 3–14.
Flickinger, Richard S., and Donley T. Studlar. 2007. “One Europe, Many Electorates?
Models of Turnout in European Parliament Elections After 2004.” Comparative
Political Studies 40 (4): 383–404.
Follesdal, Andreas, and Simon Hix. 2006. “Why There Is a Democratic Deficit in the
EU: A Response to Majone and Moravcsik.” Journal of Common Market Studies
44 (3): 533–62.
Ginsberg, Benjamin, and Robert Weissberg. 1978. “Elections and the Mobilization of
Popular Support.” American Journal of Political Science 22 (1): 31–55.
Grönlund, Kimmo, and Maija Setälä. 2007. “Political Trust, Satisfaction and Voter
Turnout.” Comparative European Politics 5 (4): 400–422.
Hadjar, Andreas, and Michael Beck. 2010. “Who Does Not Participate in Elections in
Europe and Why Is This?” European Societies 12 (4): 521–42.
Henderson, Ailsa. 2008. “Satisfaction with Democracy: The Impact of Winning and
Losing in Westminster Systems.” Journal of Elections, Public Opinion and
Parties 18 (1): 3–26.
Huntington, Samuel P. 1991. The Thrd Wave: Democratization in the Late 20th
Century. Norman: University of Oklahoma Press.
Kostadinova, Tatiana. 2003. “Voter Turnout Dynamics in post-Communist Europe.”
European Journal of Political Research 42 (6): 741–59.
Kumlin, Staffan, and Peter Esaiasson. 2012. “Scandal Fatigue? Scandal Elections and
Satisfaction with Democracy in Western Europe, 1977–2007.” British Journal
of Political Science 42 (2): 263–82.
Linde, Jonas. 2012. “Why Feed the Hand That Bites You? Perceptions of Procedural
Fairness and System Support in Post-Communist Democracies.” European
Journal of Political Research 51 (3): 410–34.
16
Linde, Jonas, and Joakim Ekman. 2003. “Satisfaction with Democracy: A Note on a
Frequently Used Indicator in Comparative Politics.” European Journal of
Political Research 42 (3): 391–408.
Lipset, Seymour Martin. 1981. Political Man: The Social Bases of Politics. Baltimore,
MD: Johns Hopkins University Press.
Nadeau, Richard, and Andre Blais. 1993. “Accepting the Election Outcome: The
Effect of Participation on Losers’ Consent.” British Journal of Political Science
23 (4): 553–63.
Norris, Pippa. 1999. “Introduction: The Growth of Critical Citizens.” In Critical
Citizens: Global Support for Democratic Governance, ed. Norris, Pippa.
Oxford: Oxford University Press, 1–27.
———. 2011. Democratic Deficit: Critical Citizens Revisited. New York: Cambridge
University Press.
Pacek, Alexander C, Grigore Pop-Eleches, and Joshua A. Tucker. 2009.
“Disenchanted or Discerning: Voter Turnout in Post-Communist Countries.”
Journal of Politics 71(2): 473–91.
Reif, Karlheinz, and Hermann Schmitt. 1980. “Nine Second-Order National Elections
– a Conceptual Framework for the Analysis of European Election Results.”
European Journal of Political Research 8 (1): 3–44.
Singh, Shane, Ekrem Karakoç, and André Blais. 2012. “Differentiating Winners: How
Elections Affect Satisfaction with Democracy.” Electoral Studies 31 (1): 201–
11.
Singh, Shane P. 2014. “Not All Election Winners Are Equal: Satisfaction with
Democracy and the Nature of the Vote.” European Journal of Political Research
53 (2): 308–27.
Singh, Shane P., and Judd R. Thornton. 2016. “Strange Bedfellows: Coalition Makeup
and Perceptions of Democratic Performance among Electoral Winners.”
Electoral Studies 42: 114–25.
The Herald. 2003. “Does a Low Election Turnout Indicate Voters Are Disillusioned
or Content?” The Herald, April 26.
Tingsten, Herbert. 1963. Political Behavior: Studies in Election Statistics. Totowa,
NJ: Bedminster Press.
Wilson, Francis G. 1936. “The Inactive Electorate and Social Revolution” The
Southwestern Social Science Quarterly 16 (4): 73–84.
17
ENDNOTES
1 We follow the conventional terminology in political science literature and consider plurality systems
as part of the majoritarian family of electoral systems.
2 British Columbia, Ontario and Quebec in Canada; Ile de France and Provence in France; Lower
Saxony in Germany; Catalonia and Madrid in Spain; Lucerne and Zürich in Switzerland.
3 The exact wording of the question was in both the pre-election and post-election surveys as follows:
“On a scale from 0 to 10 where 0 means 'not satisfied at all' and 10 means 'very satisfied,' how satisfied
are you with the way democracy works at the [election-related level of governance]?” Half of
respondents were asked this question at the beginning of each survey and the other half at the end.
4 Respondents can say that their party won, lost or that they do not know. The variable is available only
for approximately two thirds of respondents in the pooled dataset.
5 The joint inclusion of the two measurements in the regression model is not problematic since, although
they measure the same concept, they are only moderately correlated (r=0.63).
6 Full regression results are available in the Electronic Appendix.
1
Electronic Appendix to the manuscript:
The Chicken and Egg Question: Satisfaction with Democracy and Voter Turnout
Filip Kostelka
1
and André Blais
2
Table 1 Region-Election Samples Used in the Analyses
Region
Election Type
Election Year
N
Electoral System
Canada
British Columbia
National
2015
1,134
Majoritarian1
Ontario
National
2015
1,262
Majoritarian1
Ontario
Regional
2011
836
Majoritarian1
Quebec
National
2015
1,159
Majoritarian1
Quebec
Regional
2012
659
Majoritarian1
France
Ile de France
Supranational
2014
738
Proportional
Ile de France (Paris)
Municipal
2017
793
Proportional
Ile de France
National
2012
701
Majoritarian2
Provence
Supranational
2014
733
Proportional
Provence (Marseille)
Municipal
2014
481
Proportional
Provence
National
2012
676
Majoritarian2
Germany
Lower Saxony
Supranational
2014
693
Proportional
Lower Saxony
National
2013
698
Proportional3
Lower Saxony
Regional
2013
741
Proportional3
Spain
Catalonia
Supranational
2014
770
Proportional
Catalonia
National
2011
756
Proportional
Catalonia
Regional
2012
758
Proportional
Madrid
Supranational
2014
746
Proportional
Madrid
National
2011
790
Proportional
Madrid
Regional
2015
746
Proportional
Switzerland
Lucerne
National
2011
810
Proportional
Lucerne
Regional
2011
818
Proportional
Zurich
National
2011
788
Proportional
Zurich
Regional
2011
783
Proportional
Note: 1 Plurality System. 2 Two-Round System. 3 Mixed-Member Proportional System.
1
filip.kostelka@gmail.com & filip.kostelka@umontreal.ca
2
andre.blais@umontreal.ca
2
Table 2 Descriptive Statistics
Quantitative Variables
Variable Name
Count
Mean
SD
Min
Max
Pre-Election Satisfaction
19,076
0.56
0.27
0
1
Post-Election Satisfaction
19,076
0.57
0.27
0
1
Age
19,076
48.17
14.67
18
111
Interest in politics
19,076
0.67
0.26
0
1
Political Knowledge
19,076
0.68
0.35
0
1
Dummy Variables
Variable Name
Count
Frequency of yes (%)
Voted
19,706
83.3
Female
19,706
52.3
Post-Secondary Education
19,706
53.1
Duty to Vote
19,706
57.6
Feel Close to a Pol. Party
19,706
55.6
Majoritarian Electoral System
19,706
33.7
Proportional Electoral System
19,706
66.7
National Elections
19,706
46.2
Subnational Elections
19,706
34.6
Supranational Elections
19,706
19.3
Dummy Variables available for a subset of elections
Election Winner
8,463
40.7
Election Loser
8,463
43.2
Don’t know & Abstainers
8,463
16.1
3
Table 3 Regression Models Used for Figures 2-4
(1)
(2)
(3)
(4)
Election Type
(Figure 2)
Electoral System
Election Type &
Electoral
System
(Figure 3)
Election Victory
(Figure 4)
Voted
0.036 (0.008)***
0.021 (0.007)**
0.016 (0.012)
Voted * Subnational Elections
-0.009 (0.011)
0.007 (0.015)
Subnational Elections
0.005 (0.010)
0.005 (0.014)
0.041 (0.014)**
Voted * Supranational Elections
-0.039 (0.013)**
Supranational Elections 1
-0.030 (0.013)*
Voted * Majoritarian El. System
0.029 (0.012)*
0.035 (0.016)*
Majoritarian El. System 2
-0.005 (0.014)
0.001 (0.018)
0.116 (0.011)***
Voted * Subnational Elections *
Majoritarian El. System
-0.029 (0.010)**
Election Winner3
0.080 (0.008)***
Election Loser3
0.003 (0.008)
Pre-Election Satisfaction
-0.430 (0.008)***
-0.443 (0.009)***
-0.444
(0.009)***
-0.461 (0.012)***
Female
-0.007 (0.004)*
-0.007 (0.004)
-0.007 (0.004)
-0.007 (0.006)
Age
-0.000 (0.000)
0.000 (0.000)
0.000 (0.000)
0.001 (0.000)***
Post-secondary Education
0.015 (0.004)***
0.016 (0.004)***
0.015 (0.004)***
0.013 (0.006)*
Interest in Politics
0.039 (0.009)***
0.034 (0.010)***
0.034 (0.010)***
0.021 (0.013)
Political Knowledge
0.016 (0.007)*
0.021 (0.007)**
0.024 (0.008)**
0.020 (0.012)
Feel Close to a Pol. Party
0.014 (0.004)***
0.014 (0.004)**
0.014 (0.004)**
0.013 (0.006)*
Region Dummies
Yes
Yes
Yes
No
Region-Election Dummies
No
No
No
Yes
_cons
0.217 (0.012)***
0.231 (0.012)***
0.229 (0.015)***
0.104 (0.016)***
N
19076
15396
15396
8463
R2
0.25
0.26
0.26
0.30
Note: OLS Regression coefficients with robust standard errors in parentheses. * p < 0.05, ** p < 0.01,
*** p < 0.001. The dependent variable is the pre-post change in satisfaction with democracy. Model 2,
3 and 4 exclude supranational elections. 1 The reference category are national elections. 2 The reference
category are proportional electoral systems. 3 The reference category are voters who do not know
whether their preferred party won or lost, and abstainers.
4
Table 4 Replication of Table 2 without Control Variables
(1)
CS
Voted (dummy, yes)
0.014
(0.006)*
Region-Election Dummy Variables
Yes
_cons
0.002
(0.009)
N
19076
R2
0.02
Note: OLS Regression coefficients with robust standard errors in parentheses. * p < 0.05, ** p < 0.01,
*** p < 0.001. The dependent variable is the pre-post change in satisfaction with democracy.
